Gynecologic cancer remains a major health problem for women in this country with approximately 25,000 deaths annually attributed to this problem. The purpose of this project is to characterize the molecular genetics of this group of tumors and ultimately use that information for clinical application in rationally designing therapeutic and prevention trials. We have characterized endometrial, cervical, and ovarian specimens which span the histologic spectrum from benign to malignant for mutations in the ras, p53, cyclin dependent kinase inhibitors, FHIT and Rb genes and microsatellite instability. For endometrial cancers, we have analyzed curettage specimens: activated ras genes are found in the atypical hyperplasias (14%) and endometrial carcinomas (5%);p53 mutations are found in approximately 15% of atypical hyperplasia and carcinomas. Abnormalities in p15, p16, and p18 are rare in endometrial cancers. The FHIT gene is abnormal in cervical cancers and their precursor lesions but appears normal in endometrial and ovarian cancers. These studies should help to identify the molecular genetic events which are important in the genesis of endometrial and cervical cancers and their use for their early detection. Evaluation of ovarian tumors revealed that activated ras genes are found in benign (10%) and "LMP" tumors (30%) while ovarian carcinomas (5-10%) do not. In addition, mutations in the tumor suppressor genes p53 and Rb occur in ovarian carcinomas (48 and 14% respectively) but are not present in LMP tumors. This suggests that ovarian carcinoma and LMP are discrete biologic entities. In addition, abnormalities in p15, 16, 18, and microsatellite instability are extremely rare in both of these tumors. p27 expression in normal in benign and borderline tumors but decreases in malignant ovarian neoplasms. In addition, decreased expression correlates with stage of tumor. We have extended these results by examining 143 specimens of advanced ovarian cancer from a large prospective trial (GOG#111) for mutations in the p53 gene, expression of p27 and cyclin E and HER/2/neu. The results demonstrate that overexpression of cyclin E but not downregulation of p27 is a poor prognostic factor in advanced ovarian cancer. Future projects will examine the value of p53 mutations as a prognostic or an early detection marker in ovarian cancers by examining large numbers of early stage ovarian cancers, specimens from second look operations, and PAP smears from women with ovarian cancer. Finally, to identify potential new markers of ovarian cancer and genes important in its pathogenesis, malignant ovarian epithelium is being compared to its benign counterpart using differential display technology, representational display, and microarrays. As part of the Director's Challenge collaborative grant between this laboratory and MSK, 400 ovarian cancer specimens will be profiled utilizing cDNA microarrays and patterns will be correlated with clinical characteristics such as survival, histology , and stage.Genes which are differentially expressed will be isolated, cloned and characterized for their role in the development of ovarian cancer. We are presently planning a prospective phase III trial to validate the prognostic value of these genes.